he Binary Neutron Star Merger
9.2 Chronology ou tht Follow-up
Direct gravitational wave detections have always been followed by numerous multi-messenger searches, however none of them has been successful with binary black holes. For this irst bi-nary neutron star, an electromagnetic signal has been detected before any alert was sent by the gravitational wave detectors and it has been followed by many moreα[21]. More than 70 ob-servatories represented on igureα9.5took part to the multi-messenger follow-up and the paper
Figurt 9.5–Map ou tht gravitational wavt stttrtors (ytllow), light-basts obstrvatorits (blut) ans ntutrino ttltsropts (rts)that took part to the detection or follow-up of the GW170817 gravitational wave event.
reporting these has been signed by ∼3 500 scientists.
his section will present the chronology of detection of multi-messenger signals coming from the binary neutron star merger GW170817. A participation in this follow-up will also be introduced in this chronology. In order to make the reading more pleasant, only the irst detections of each messenger will be stated, however the author does not grant less scientiic interest to subsequent detections.
Most of the detections stated below are displayed on igureα9.7 with the corresponding Gamma-ray Coordinates Network ⁸GCN⁹ notices and circulars. he GCN is the network used by astronomers and astrophysicists to distribute informations about -ray bursts and transient events. A notice is an alert sent without any humans-in-the-loop while a circular is a prose-style message from follow-up observers reporting on their results. he upper part of igureα9.7 represents the real time detections while the lower part represents the informations sent to the community.
On August 17th, 2017, although the gravitational wave signal happened before, theFermi Gamma-ray Burst Monitor ⁸Fermi-GBM⁹ was the irst experiment to send a public alert to the communityα[130] at 12:41:20 UTC, 14 s ater the detection. he event, shown in igureα9.6, was seen with a 4.8 σ signiicance and a rough localization of the event with a 3 200 degl credible region.
In the meantime, the low-latency binary-coalescence search was identifying the grav-itational wave signal in the aLIGO-Handford data. he detected coalescence inished at t = 12:41:04 UTC and at this time the Livingston detector was saturated by a glitch, as for aVirgo, its low-latency data transfer was delayed. Nevertheless, a GCN notice has been released 27 min ater the merger and 13 min later the GCN circular Nt 21505α[131] linked this event with the -ray burst seen byFermi-GBM at t ⁻ 1.7 s. he atempt to localize the event with one detector
Figurt 9.6–Light rurvt ou GaB 170817Ameasured byFermi-GBM in the 50 to 300 keV band. he red band is the background estimate. α[20]
9.heBinaryNeutronStarMerger
Figurt 9.7–cimtlint ou tht multi-mtsstngtr obstrvationsof the GW170817 event ⁸green⁹ as well as the Gamma-ray Coordinates Network
⁸GCN⁹ notices ⁸blue⁹ and circulars ⁸red⁹ sent subsequently in the 18 hr following the event. he A GCN circular is represented in orange.
Figurt 9.8– 3′ × 3′ imagts rtntrts on NGC 4993 with North up ans East ltt. Pantl A:Hubble Space Telescope image from four months before the merger. Pantl B:Swope image of the binary neutron star remnant denoted as SSS17a on August 17th, 2017 at 23:33 UTC. SSS17a is marked with the red arrowα[135].
lead to a credible region much broader than theFermi-GBM one.
It is from these measurements that A started a search for a neutrino counterpart using the online sample of upward-going tracks. At this time the credible region of Fermi-GBM and aLIGO-aVirgo were partly in the A ield of view. However, in parallel LIGO Scientiic Collaboration removed the background from the aLIGO-Livingston detector and Virgo Collaboration processed their data. With the data of the three interferometers, the credible region was reduced to 31 deglα[132]. As a consequence, the localization of GW170817 was not in the A ield of view any more when the results of the online analysis were sentα[133], eight hours ater the merger detection. No neutrino event were seen either in a time window of k 500 s around the coalescence time t nor in a more extended time window of k 1 hr. his analysis will be developed in sectionα10.1.2.
he optical telescopes strategy was to target galaxies inside of the three-dimensional local-ization of the event accounting for their stellar mass and star formation rate. he locallocal-ization using the three gravitational wave detectors reduced a lot the number of targeted galaxies but the credible region was not in the ield of view of the terrestrial telescopes before ten hours and the Chilean night. he spatial telescopes have a narrower ield of view and did not follow this event before it was well localized. It was the 1 m Swope Telescope that irst detected light at 23:33 UTCα[134] and located the event in the galaxy NGCα4993 as can be seen in igureα9.8. Five other optical detections followed within an hour.
he Rapid Eye Mount/ROS2 detected the irst near-infrared signal 12.7 hours ater the coa-lescence followed by ultraviolet detection by the UVOT instrument on-boardSwit satellite at t ⁻ 15.3 hr. hese observations were continuous during the following days and showed an un-usual rapid luminosity decline in UV-blue and brightening of the near-infrared emission. his
evolution, characteristic of a kilonova, is considered unprecedented by the community for a transient event in nearby universe and makes this event even more interesting.
At t ⁻ 4αdays, A sent the GCN circular Nt 21631α[136] reporting the results of a search for downward-going tracks, as the merger was in the downward-going ield of view at� . As developed in sectionα10.2.1, no event spatially correlated with the merger passed the cuts neither in the k 500 s nor in the k 1 hr time windows.
X-ray and radio observations are very useful to constrain the geometry of the ejecta, energy output as well as the orientation of the system and the environment of the merger. During the irst days neither X-ray nor radio emissions had been detected and limits were put. It is only nine days ater the event that the irst X-ray counterpart was detected byChandra and seven days later for the irst radio counterpart with the Jansky Very Large Array.
hese multi-messenger searches have important physical implications that will be developed in the following section.
In parallel, A started two oline analyses onto which I have contributed by adding sensitivity to all the neutrino lavour events using the shower sample on a k 500 s time window around the merger time and ⁻ 14αdays ater the coalescence. hese analyses will be developed in sectionsα10.2.2andα10.2.3.